Merlion—ARIMA时间序列预测金融数据

数据样式

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from merlion.utils import TimeSeries
from merlion.models.forecast.arima import Arima, ArimaConfig
from merlion.transform.base import Identity
from merlion.evaluate.forecast import ForecastMetric

# 参数
max_forecast_steps  = 100
target_seq_index = 0
order=(25, 1, 6)

def read_data():
    """
    读取数据
    """
    df = pd.read_csv('Predict_Data_6.csv')
    df['month'] = pd.to_datetime(df['month'])
    df = df.set_index('month')
    return df

def df_to_timeseries(df):
    """
    转换格式，划分数据集
    """
    time_series_train = TimeSeries.from_pd(df.iloc[:int(df.shape[0]*0.8),:])
    time_series_test = TimeSeries.from_pd(df.iloc[int(df.shape[0]*0.8):,:])
    return time_series_train,time_series_test

def get_target(test):
    """
    获取需要预测的目标变量真值
    """
    target_univariate = test.univariates[test.names[target_seq_index]]
    target = target_univariate[:max_forecast_steps].to_ts()
    return target

def build_model():
    """
    构造模型
    """
    config1 = ArimaConfig(max_forecast_steps=max_forecast_steps, target_seq_index=target_seq_index, order=order,
                          transform=Identity())
    model1 = Arima(config1)
    return model1

def assessment(forecast):
    pred = forecast.univariates['closeLogDiff']
    test_df = df.iloc[int(df.shape[0]*0.8):,:]
    test_df['forcast'] = pred
    test_df = test_df.drop(labels=['diff(G1000037)_9_abs','diff(G0008003)_12_abs','diff(G0000029)_12','logDiff(G0001596)_12','P9918147_exp','diff(G1400003)_12_exp'],axis=1)
    test_df['true_tend']=np.where(test_df.closeLogDiff.diff(periods=1) >= 0, 1, 0)
    test_df['pred_tend']=np.where(test_df.forcast.diff(periods=1) >= 0, 1, 0)
    test_df['pred_tend'][0]=1
#     accuracy = accuracy_score(test_df['true_tend'],test_df['pred_tend'])
#     print(f'相同趋势准确率{accuracy:.4f}')
    a = len(test_df[test_df['closeLogDiff']>0][test_df['forcast']>0])
    b = len(test_df[test_df['closeLogDiff']<0][test_df['forcast']<0])
    accuracy = (a+b)/len(test_df)
    print(f'相同趋势准确率{accuracy:.4f}')
    return test_df,accuracy

if __name__=='__main__':
    df = read_data()
    train,test = df_to_timeseries(df)
    target = get_target(test)
    model1 = build_model()

    # 训练模型
    for model in [model1]:
        print(f"Training {type(model).__name__}...")
        train_pred, train_stderr = model.train(train)

    # 进行预测
    for model in [model1]:
        forecast, stderr = model.forecast(target.time_stamps)
        rmse = ForecastMetric.RMSE.value(ground_truth=target, predict=forecast)
        smape = ForecastMetric.sMAPE.value(ground_truth=target, predict=forecast)
        print(f"{type(model).__name__}")
        print(f"RMSE:  {rmse:.4f}")
        print(f"sMAPE: {smape:.4f}")
        print()

    # 可视化
    fig, ax = model.plot_forecast(time_series=test, plot_forecast_uncertainty=True)
    plt.show()
#    plt.savefig('Arima预测')

    result_data, accuracy = assessment(forecast, df)

Arima
RMSE: 0.0355
sMAPE: 115.4556